Off by default (Statistics:Crawler:Enabled), as the owner decided. When it is on:
- CrawlPlan runs hourly, from StatisticsSchedule. It inserts the configured seeds and
queues up to HostsPerHour servers whose last visit is older than RevisitDays, are not
paused by the breaker and are not domain-blocked, spread across the hour.
- CrawlInstance visits one server at a time as
"PrivaPub-Stargazer/<ref> (+<base>/stargazing)":
- it reads robots.txt (RFC 9309: its own group first, then PrivaPub, then *; longest
rule wins; a 4xx allows everything; a 5xx or no answer keeps it out);
- it describes servers that only crawling ever found, through
InstanceDescriber.Describe with robots.txt as the path filter;
- it reads /api/v1/instance/peers through the new GetStringArray, which keeps what it
read from the first 1 MB instead of refusing a large list;
- it adds the names a server could ever be reached at as "crawled": DNS only,
punycode, no addresses, ports or hidden services, and the reserved test names only on
a test network. Never more than MaxNewHostsPerCrawl per visit or MaxHosts in all, and
never over a touched server.
- It reads nothing but robots.txt, NodeInfo, the instance API and the peers list.
IFederationHttp.GetText serves robots.txt, and HttpScope.Crawl carries the
User-Agent.
- /stargazing explains all this and how to keep the crawler out, says whether it is on,
and credits DB-IP. GET /clientapi/admin/statistics/crawler shows the frontier.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01ELjqpznMFMNrJoJUj6K5p2
Communities:
- a post addressed to a community (to, cc or audience) is accepted
according to its posting policy - followers, anyone, or moderators -
and GroupDistributor announces the whole activity with `audience` to
the community's followers, plus the object for new posts so Mastodon
shows them; updates and deletes of community content are announced too;
- top-level posts are Pages with a name (the title, or a headline from
the text); /flock counts members, /wardens lists moderators;
- a Mastodon client posts into a community by mentioning it, or into a
remote group, which sets `audience`;
- an Announce of an activity from a remote group a persona follows (Lemmy)
is followed through: the object is fetched from its own origin, kept with
its AudienceURI, and fanned out to the group's local followers; updates
are applied in place and deletes checked against the origin.
Circles stop being local-only: an undiscoverable Group actor whose follows
are all requests the owner approves; posts addressed to the circle and its
/flock and delivered to members' own inboxes, never announced, never
public; a remote member's post into the circle is accepted from members
only. SignedFetchAuthorizer serves circle posts and collections only to a
signed request from a member or a member server's instance actor - 404 for
anyone else. Circles never surface in search, lookups, mentions, account
ids or profile pages.
Federation:SecureMode requires a valid signature on every GET under
/peasants except the instance actor. Group forms take a posting policy.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
Domain/Statuses/StatusService takes a persona, not a root, and is what
/clientapi and the Mastodon API share:
- Publish renders Markdown (clientapi) or plain text (Mastodon clients),
checks the persona may see the post it replies to, opens or reuses the
conversation for a direct post from its recipients and mentions, fans
out and hands the Create to the outbox;
- Edit keeps a revision and sends Update{Note}; Remove soft-deletes and
returns the post, so a client can delete and redraft;
- Favourite and Reblog work on anything the persona can see and send Like,
Announce and their Undo to the author (and, for boosts, to followers);
boosts are refused for anything but public and unlisted posts.
PostsService is now the clientapi wrapper that checks the root owns the
persona. A persona's own boost is a local row: the outbox renders it as
Announce, /grunts/announce-{id} resolves, and object endpoints, profile
pages and NodeInfo skip it. ContentFormat gains Plain.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
VisibilityPolicy.IsPublic is the single expression for "anyone may see
this" (Public or Unlisted, not deleted); the outbox, the object and
activity endpoints, the HTML pages and NodeInfo all use it instead of
spelling it out. CanSee answers for a persona: the author, a mentioned
local persona, a conversation member for Direct, a circle member for
Circle; followers-only waits for P1.2's follows.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
/@user and /@user/{id} are Razor pages showing an avatar's or community's
public and unlisted posts: no scripts, a strict CSP, no-referrer, noindex,
and an alternate link to the ActivityPub document. A client asking them
for activity+json is redirected to the actor or note, and the actor and
note redirect browsers here.
NodeInfo answers 2.1 as well as 2.0 (repository, homepage, a link to
FEDERATION.md) and counts only public local posts.
FEDERATION.md (FEP-67ff) lists the protocols, FEPs, route names,
activities and the security rules a peer will notice.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012CzABvBkbcFqoHdmi8b9WB
thepraandClaude Opus 5.5